I am a slow learner, and I can’t remember and learn algorithms as well as my friends do. I can’t even implement SCC on my own! But that doesn’t matter; Codeforces nowadays doesn’t require almost any advanced algorithms in <3000 rated problems. It might sound surprising to some beginners, but knowing BFS, DFS, union-find, Dijkstra, MST, hashing, inverse elements, gcd caluculation, fast nCk calculation algorithm, and segment trees are enough to reach red! So, learning advanced algorithms won’t help much when trying to reach red. Fundamental problem-solving skills are way more important.

When practicing, always pick problems hard enough to challenge you but easy enough, so you have a chance to solve them without editorial. This was my mistake for a long time; I kept picking easy problems because of the satisfaction of getting AC. Then I realized the hard way (finishing almost last at OI TST) that I won’t improve like this. So what did I do? I started picking actual hard problems but I always failed to solve them and quickly jumped to editorial. Now, this wasn’t all bad; I learned a few advanced problem-solving techniques like this, which helped me better understand and implement complicated ideas. But honestly, I was wasting time practicing inefficiently. So right now, I try to pick myself challenging problems (in the range around CF 2500-2800, but that’s not very important) such that I need to think about them for a few hours until I figure out the solution. I almost always do them on my own without reading editorials. And this is the part of CP I like most. There’s nothing as satisfying as finally figuring out a solution to a hard problem you couldn’t solve for hours and then getting AC.

Always implement the problem after reading the editorial/solving it, don’t just settle for “mind solving” it. Implementation is an essential part of competitive programming and shouldn’t be underestimated.

It is crucial to build intuition about problem-solving. I am not done with the problem when I get AC while practicing. I always think about it. How did I solve it? Did I learn something new? What part of the problem was most problematic for me? Is there anything I could learn for the future from this problem? I have a diary where I write down answers to such questions after solving every problem. I started doing this only recently, but I think it helped me build better problem-solving intuition.

Have fun when practicing! Don’t take CP or rating too seriously. Find a way of practicing you can enjoy. Compete with friends, challenge each other, and discuss problems. I underestimated this a lot in my early years. But it is perhaps the most crucial aspect of training.

Always stay calm during the contest. Focus on solving the following problem and nothing else. Don’t think about how much rating you will win or lose, don’t think about beating a friend. For now, there’s only the next problem ahead of you, which you must solve. When you solve it, you win. You can do it. And even if you fail to solve it, you will learn something and fix your rating soon.

Only start implementing after you are 100% sure you know how to solve the problem or there’s little time left. Otherwise, you might lose precious time implementing something wrong or get many wrong answers.

Check the dashboard every 15 minutes or so to see the solved counts and get an idea of how hard the following problems are.

When you get WA/TLE/RE and don’t know why, reread your code. Check the constraints in the problem. Try to run it on a few small test cases, which you can find an answer to using hand. Try to run it on the max test case on CF custom invocation. Tip: Use Clang++20 to check for overflows/out-of-bound issues and stuff. If you still don’t know what to do, stress test. It’s easy if you practice it a lot. It can take less than 20 minutes even.

Never quit the contest too early. Always at least read and attempt the following problem. There’s always a chance for you to solve it.